Know your cognitive environment! Mental models as crucial determinant of offloading preferences
Patrick P. Weis, Eva Wiese · 2019
Objective: Human problem solvers possess the ability to outsource parts of their mental pro-cessing onto cognitive “helpers” (cognitive offloading). However, suboptimal decisions regarding which helper to recruit for which task occur frequently. Here, we investigate if understanding and adjusting a specific subcomponent of mental models –beliefs about task-specific expertise – regarding these helpers could provide a comparatively easy way to improve offloading decisions. Background: Mental models afford storage of beliefs about a helper that can be retrieved when needed. Methods: Arithmetic and social problems were solved by 192 participants. Participants could – in addition to solving a task on their own – offload cognitive processing onto a human, a robot, or one of two smartphone apps. These helpers were introduced with either task-specific (e.g., stating that an app would use machine learning to “recognize faces” and “read emotions”) or task-unspecific (e.g., stating that an app was built for solving “complex cognitive tasks”) descriptions of their expertise. Results: Providing task-specific expertise information heavily altered offloading behavior for apps but much less so for humans or robots. This suggests 1) strong pre-existing mental models of human and robot helpers and 2) a strong impact of mental model adjustment for novel helpers like unfamiliar smartphone apps. Conclusion: Creating and refining mental models is an easy approach to adjust offloading preferences and thus improve interactions with cognitive environments. Application: To efficiently work in environments in which problem solving includes consulting other people or cognitive tools (“helpers”), accurate mental models –especially regarding task-relevant expertise– are a crucial prerequisite.